ABSTRACT Bathymetric lake mappers are inconsistently employing survey techniques and GPS/Sonar equipment. Comparing transect spacing and GPS/Sonar equipment for fluvial lake mapping is needed. A bathymetric survey was conducted on the Stevens Point Flowage in 2013 from Bukolt Park north to the Highway 10 overpass. The survey was completed with two methods including 1) a Trimble R6 GPS receiver, TSC2 data collector, Sonarmite depth sounder, cellular device with WISCORS instantaneous corrections applied and 2) a Lowrance HDS5 GPS/Sonar. Transects were initially spaced 300 feet apart running perpendicular to flow direction. Trimble and Lowrance XYZ points that occurred within 30 feet of each other were used to create separate lake models for each technique. Subsequent models were built after manually removing every other transect (600 foot spacing). Flowage models from both survey techniques and varying transect spacing scenarios were built and compared at randomly generated locations (two-way, paired t-Test(two-tailed)317 using ESRI ArcGIS 10.1 software and the Geospatial Modeling Environment tools to determine if significant differences exist between survey techniques. Pool volume differences are also evaluated.

STUDY AREA The Stevens Point flowage is an impoundment of the Wisconsin River located in Portage County (Figure 1). The surveyed area extends from Bukolt Park north to the Highway 10 overpass. The studied area is 2611 acres with a channel length of seven miles.

Comparison of transect spacing and GPS/Sonar equipment for fluvial lake mapping

Author: Mason Johnson Advisor: Christine Koeller

INTRODUCTION Surveyors are able to measure lake depth fairly well using modern sonars but, these types of measurements can be challenging if striving for the highest level of accuracy (Fridl 2009). Utilizing echo sounder and GPS technology to estimate lake volume requires precision for developing a representative model of the system. Bathymetric data, in essence information about the water depth and underwater topography of oceans, seas and lakes, are important in many aspects of marine and lacustrine research, administration and spatial planning of marine and coastal environments and their resources (Jakobsson 2011).

DISCUSSION High cost, survey-grade equipment (Trimble), produces accurate GPS locations to the sub-centimeter level. Low-cost survey equipment (Lowrance) failed to produce an accurate representation of the Stevens Point Flowage when compared to Trimble. Error associated with Lowrance survey equipment resides in the inability to receive instantaneous coordinate corrections (Figure 9). Modeling lakes require a higher level of accuracy to represent them, depending on the objective. Transects spaced 300 feet apart result in more accurate lake volume estimations. Increasing transect spacing will likely reduce accurate lake volume estimations. This poses a problem when developing best management practices. Lake managers controlling aquatic invasive species often use water level drawdown techniques to eliminate these species. Drawing-down lakes to an extreme level can result in native plant’s being unable to establish. Knowing lake volume reduces this risk.

OBJECTIVES 1) Determine if significant differences exist between survey-grade (Trimble R6) and mapping-grade equipment

(Lowrance HDS5).

2) Determine adequate transect spacing using survey-grade equipment for estimating lake volume. (µ T300 = µ T600)

Figure 2: Trimble/Lowrance survey equipment attached to transform of 6hp Jon boat while surveying .

METHODS Data Acquisition for 2013 Maps

• GPS/Sonar survey in summer-fall 2013 using survey-grade equipment (Trimble R6 GPS, Ohmex Sonarmite depth sounder) (Figure 2).

• Real-time corrections were applied to Trimble equipment via WISCORS reference station (<1 cm horizontally, ±3cm vertically).

• Grid transects were spaced 300 feet apart perpendicular to channel flow (Figure 3). • Near shore depths were also acquired. • XYZ positions were recorded every second. • Lowrance survey data was collected using a similar method devoid of WISCORS differential corrections.

Figure 3: Grid-like transects spaced 300 feet apart and proximate shore outline, Stevens Point Flowage spring 2013.

Figure 7: Trimble 1m raster. Figure 6: Lowrance 1m raster.

Figure 10: Random points generated within each 10-acre grid.

References:

Aquatic Plant Management - Water Level Drawdown. Available at http://www.ecy.wa.gov/programs/wq/plants/management/drawdown.html (verified30 March 2014).

Beyer, H.L. (2012). Geospatial Modelling Environment (Version 0.7.2.0). (software). URL: http://www.spatialecology.com/gme.

Fridl, J., Kolega, N., & Žerjal, A. (2009). Applications of Seafloor Mapping Using Precise Sonars. Cartography & Geoinformation, 8(11), 14-24.

Hell, B., Broman, B., Jakobsson, L., Jakobsson, M., Magnusson, Å., & Wiberg, P. (2012). The use of bathymetric data in society and science: A review from the baltic sea. Ambio, 41(2), 138-50. doi:http://dx.doi.org/10.1007/s13280-011-0192-y.

Acknowledgments:

Christine Koeller, Advisor.

Doug Miskowiak, GIS Education Specialist.

Prof. Keith Rice, GIS Center Director.

UWSP GIS Center for use of survey equipment.

Figure 4: Transects spaced 300 feet manually removed.

RESULTS Table 1:Percent difference of lake volumes produced by Trimble vs Lowrance depth models.

Table 3: Paired t-Test of Lowrance and Trimble bathymetric models at random points within ten-acre grid.

Table 4: Paired t-Test of Trimble bathymetric models using 300 foot transects.

Paired t-Test (α=0.05) t Critical P-value

Trimble t(two-tail),317 -2.68 0.008

Paired t-Test (α=0.05) t Critical P-value

Lowrance vs. Trimble t(two-tail),317 -9.94 0

Lake Volume (Acre-feet)

Trimble 10254.22

Lowrance 9297.64

Average 9775.93

Difference 956.58

Percent Difference 9.79

Transect Spacing (feet) Volume (Acre-feet)

300 10254.22

600 9894.37

Average 10074.30

Difference 359.85

Percent Difference 3.57

Table 2: Calculated percent difference of Trimble lake volumes with different transect spacing.

Significant difference exist (P<0.05) between Lowrance and Trimble survey equipment.

Significant difference exist (P<0.05) between Trimble 300 and 600 foot transect spacing.

Figure 8: Depth variations of Lowrance model compared to Trimble

Graph 1: Depth distribution of XYZ locations on Stevens Point Flowage, 2013.

Figure 9: Trimble 1m raster grid with 16x exaggerated heights.

Figure 5: Positional and depth error of Lowrance survey data.

Figure 1: Stevens Point Flowage, Portage County, WI.

METHODS Model Comparisons

• Utilizing Geospatial Modeling Environment, one-acre grids were created (Figure 10). • Random points were generated for each ten-acre grid. • Depth points were extracted from each 1m raster model using generated random points. • Paired t-Test executed to determine if any significant differences existed between Trimble and Lowrance (Table 1 and 2). • Percent differences were calculated for survey equipment (Table 3) and Trimble transect spacing (Table 4).

METHODS Construction of 2013 Trimble and Lowrance Surface Models

• Shoreline and islands were digitized from WROC 2010 aerial imagery using ArcMap 10.2 software. • Vertices from digitized shoreline and islands were set to a depth of zero and combined with surveyed data. • Transects spaced 300 feet apart were manually removed for comparison between 300 and 600 foot models (Figure 4). • Triangulated Irregular Network (TIN) models were created from Trimble and Lowrance data . • TIN models were converted to 1m raster grid (Figures 6 and 7).

Difference in triangulated surfaces.